Thoughts on whether the model can be biased?

[formal elk]

Currently doing some research on the social and cultural dimensions of technology and am interesting in hearing peoples thoughts on this.

[upbeat vigil]

I don't think it's really possible to be unbiased in a conversational context. the things you choose to say are just as important as the things you choose to leave out, and the LLM leaving nothing out would make it just... not useful